To improve the production and transportation efficiency of the automatic material handling system,a graph network model of tracks was established according to the hybrid layout,and a dynamic scheduling policy based on reinforcement learning was proposed.The scheduling process consisted of two stages,the transportation task assignment and the transportation path planning.The task assignment was solved using Hungary algorithm based on real-time traffic information.The transportation path planning problem was formulated as the path decision problem.A reinforcement learning model based on Q-Learning(off policy)and Boltzmann exploration strategy was used to solve path decision problem.The final experimental results show that the proposed policy can improve the production and transportation efficiency of the automatic material handling system.
关键词
自动化物料运输系统/高空提升搬运车/调度/路径决策/强化学习/任务指派/路径规划/交通信息
Key words
automatic material handling system/overhead hoist transport/scheduling/path decision/reinforcement learning/task assignment/path planning/traffic information
维普
万方数据